evidence-first

Give your AI agent eyes for PDFs — structured text, tables, OCR, visual evidence, and page-level citations via MCP. Native Rust, local-first.

AGI-oriented, hallucination-resistant AI coding runtime grounded in evidence, tools, memory, agents, and verification.

Evidence-first runtime for a wheeled mobile household robot with bounded actions, replayable events, and fail-closed verification.

Free-first Chinese and English web search MCP using zero-key sources and inspectable evidence.

Mini Program Engineering Skill Suite is an Agent Skill suite for evidence‑first mini‑program development. It helps agents bring WeChat or other mini‑program projects from vague intent to reliable engineering work: project intake, product specification, architecture, implementation, UI & device adaptation, debugging, verification, release readiness.

数学建模竞赛全流程 AI Skill：以证据优先的 mmflow 运行时为唯一事实来源，12 阶段状态机（P0–P11）逐阶段 fail-closed 门禁，哈希链账本不可变、结果可复现、隐私扫描、heal_mismatch 规范恢复。38 个 CLI 命令、56 个模板（7 类目）、15 份手册、46 种证据类型；LaTeX 论文真实编译、退化测试与反证、多角色独立评审闭环。23 项契约校验、237 个测试函数、端到端全流程验证 COMPLETE。支持 CUMCM、MCM/ICM、GMCM、MathorCup、APMCM、深圳杯；适配 Claude Code/Codex 等 Agent Skills。MIT 开源。

An accountability layer for investment agents: thesis cards, prediction ledgers, stock screeners, and feedback loops.

Evidence-first 1C:ERP diagnostics — dynamic Gate 0–10 orchestration and a verified Unica/1C Skills companion ecosystem.

Execution harness for OpenClaw agents: evidence-first changes, verification gates, dangerous command controls, and codex-harness / claude-code-harness discoverability.

Evidence-first workflow protocol for vehicle diagnostic AI agents.

Evidence-first coding agent for Claude Code CLI. Port of burkeholland/anvil from GitHub Copilot CLI.

Evidence-first workflow protocol for vehicle diagnostic AI agents.

Legal AI research, evidence-first engineering, AI safety, REG-containers, and responsible legal technology by EN-DO Technology.

v0.2 Beta: evidence-first local research workflow skills with audited MCP writes for Chinese undergrad information-management empirical theses.

Signed, sanitized records of real defensive security work - every artifact ships with a SHA-256 sidecar for independent verification.

Evidence-first AI search for public Peruvian cultural and educational resources.

Evidence-first incident diagnosis. Root cause in 16–30s. Project 1.0.0 complete (29/29 checkpoints).

Evidence-first video understanding skill for Codex: yt-dlp, Douyin SSR fallback, local whisper.cpp, keyframes, OCR, and grounded reports.

1337 Security Workbench — open security workstation and model-agnostic security execution, state, and evidence platform

Evidence-first, human-governed job-application operations: traceable career evidence, role-specific documents, ATS preparation, and outcome records.